ABSTRACT
Linear models always have enjoyed a special status in engineering, despite nature itself being mostly nonlinear. The use of linear models was paramount before the days of digital computers, because usually only they possess analytical solution. Much more important today is the fact that the behavior of linear models can be characterized in vastly simpler ways than almost all nonlinear systems, regardless of whether mathematics is employed. The significance of this fact has not waned; well engineered systems ought to behave in simple, predictable ways, so linear behavior remains a frequent object of design. Third, strategies for the control of dynamic systems predominantly depend on the use of at least approximate linear models. Therefore, even nonlinear models are often linearized, despite the approximateness of the result. Finally, systems in which the variables range over only a small fraction of that which is possible, such as with most acoustics and other vibrations, are very nearly linear anyway.
